Here are two more options for viewing this letter that describes trends
in publication speed for Journal of the Atmospheric Sciences between
1970 and 1997. The BAMS article was declined by the AMS. I think it
was viewed as an attack on the AMS leadership or journal editors, but I
was just trying to raise an issue for which I think we all have some
responsibility. Results for year 2000 suggest the situation is getting worse.
September 15, 2000: . I recently heard that the slowdown in 2000 was the result of problems with the publishing house, and that we can hope the situation will improve somewhat in the future, perhaps returning us to the still not ideal situation of 1997.
1) PDF version of letter
2) PDF of paper submitted to BAMS 6/25/97
3) Postscript version of paper
It looks like the review process is holding steady at about 8.5 months, but the AMS publication process has slipped out to 10.7 months for a total time of about 19 months for first submission to publication date, compared to 15.4 months for the larger sample I used in 1997. Why does it take AMS 11 months to get refereed papers into print? At this rate of increase, by 2004 it will take two years between first submission and publication date. So if you submit a paper on January 1 of 2002, it will appear in about the January 1, 2004 issue. I hope this extrapolation is inaccurate. Average page length is holding steady at about 16.7 compared to 18 in my 1997 sample.
I did the same analysis for the January 2000 issue of Journal of Climate. The numbers are remarkably similar.
Sample size is 18 papers with average length 16.7 pages.
The HTML version of the 1997 letter is below.
Department of Atmospheric Sciences, Box
351640 . . . . . . . . . . . . . . (206) 543-7460
FAX: (206) 685-9302 . . . . . . . . . . . . . .
dennis@atmos.washington.edu
10 June 1997
Anyone Concerned
Dear Colleague:
Annoyed at the length of time it takes to get a review back from JAS, I did a little research project to see how long it takes for a paper to get into print. I compared the first three months of JAS in 1997 (6 issues for Jan., Feb. Mar.) with the first three months in 1970 (Jan., Mar. May). I used the publication date on the journal to measure the publication time. It is a small sample of about 80 papers, but I think the conclusions are robust. Numbers for other AMS journals are probably similar. I have not included Notes and Correspondence in the calculations of page length or publication delays. Here are the results:
1970 . . . . 44 . . . . . 10.4 . . . . . . . 20 . . . . . . . . . 179 . . . . . . . . 134 . . . . . . . . . .45 . . . . . . . . . 20
1997 . . . . 39 . . . . . 18.2 . . . . . . . . 7 . . . . . . . . . 461 . . . . . . . . 202 . . . . . . . . 260 . . . . . . . . . . 0
Codes: S-P = Submission to Publication
Delay, A-P = Acceptance to Publication Delay,
S-A = Submission to Acceptance Delay. All in days.
1) Average time between first submission and publication has increased from 6 months to 15 months.
2) Of the additional 9 months that it now takes new work to appear in print, 7 are accounted for by the review process and 2 are accounted for by the publication process. On average, it now takes more than 8 months to complete the review process and another 7 months after acceptance before publication occurs. This is to be compared with six weeks and 4.5 months in 1970.
3) Articles have nearly doubled in average length since 1970 from 10 pages to 18.
4) Much less use is made of Notes and Correspondence
5) In 1970, almost half the papers were published as first submitted. This never happens now.
6) All of the above is troubling, especially 1 and 2. In terms of scientific communcation to foster growth of the knowledge base, JAS is now much less effective than it was in 1970, because of the large delay between when work is current and when it is published.
Sincerely,
Dennis L. Hartmann
Professor
Back to Dennis Hartmann's Home Page